Claude aids Iran strikes as Trump bans Anthropic

The core irony is stunning: the Pentagon was deploying Claude for targeting while Trump signed the order banning Anthropic from federal contracts.

• –Claude was used for "intelligence assessments, target identification, and simulating battle scenarios" — framed as decision support, not autonomous weapons control, but the line is blurry
• –The dispute originated from Anthropic's push for explicit guardrails against mass surveillance of Americans and fully autonomous weapons; the Pentagon insisted on unrestricted access for "all lawful purposes"
• –Trump's ban gives agencies six months to transition, but Defense One reports replacing Claude's capabilities could take three months or more — suggesting the ban is partly performative
• –If AI models can be deployed for lethal targeting without maker consent to specific use terms, the concept of enforceable AI "red lines" becomes nearly meaningless
• –For developers and AI labs pursuing federal contracts, this signals that government deals will increasingly require forfeiting safety restrictions — a tension every lab with DoD ambitions must now confront